.env.example

# Economic Data Project - Environment Variables
# Copy this file to .env and fill in your values

# =============================================================================
# ENVIRONMENT
# =============================================================================
ENVIRONMENT=dev  # dev or prod

# =============================================================================
# ENVIRONMENT (controls which BigQuery dataset suffix is used)
# =============================================================================
# prod    → economics_raw, economics_staging, economics_marts, ... (no suffix)
# staging → economics_raw_staging, economics_staging_staging, ...
# dev     → economics_raw_dev, economics_staging_dev, ...
# All three environments live in the same GCP project.
ENVIRONMENT=dev

# =============================================================================
# DATABASE - BigQuery (primary warehouse, all environments)
# =============================================================================
BIGQUERY_PROJECT=your-gcp-project-id
# BIGQUERY_DATASET is auto-derived from ENVIRONMENT (economics_raw_dev by default).
# Override only if you need a custom dataset name.
# BIGQUERY_DATASET=economics_raw_dev
BIGQUERY_LOCATION=US
ICEBERG_BUCKET_NAME=econ-project-iceberg-data

# Metaxy metadata store (issue #46). Connection string for the Metaxy
# lineage/versioning catalog (BigQuery project).
METAXY_STORES__PROD__CONFIG__DATABASE=

# =============================================================================
# DATA SOURCE APIs
# =============================================================================
FRED_API_KEY=your_fred_api_key_here
MARKETSTACK_API_KEY=your_marketstack_api_key_here
CENSUS_API_KEY=your_census_api_key_here

# =============================================================================
# SEC EDGAR API
# =============================================================================
# Contact email required by SEC for API access (User-Agent header)
# SEC requires identification for rate limiting and abuse prevention
# Use your real email - SEC may contact you if there are issues
SEC_EDGAR_CONTACT_EMAIL=your_email@example.com

# =============================================================================
# LLM / AI PROVIDERS
# =============================================================================
LLM_PROVIDER=openai  # openai, anthropic, or gemini
MODEL_NAME=gpt-4-turbo-preview

# OpenAI
OPENAI_API_KEY=your_openai_api_key_here

# Anthropic (optional)
ANTHROPIC_API_KEY=your_anthropic_api_key_here

# Google Gemini (optional)
GEMINI_API_KEY=your_gemini_api_key_here

# =============================================================================
# GOOGLE CLOUD (for GCS storage, Google Docs, and BigQuery)
# =============================================================================
# Use Application Default Credentials (ADC) for Google auth.
# Local setup:
#   gcloud auth application-default login
# Service account impersonation, if preferred:
#   gcloud auth application-default login --impersonate-service-account service-account@project.iam.gserviceaccount.com
# On GCE/Cloud Run/GitHub WIF, ADC uses the attached or federated service account.
#
# GOOGLE_APPLICATION_CREDENTIALS is intentionally unset for normal ADC auth.
# Set it only if you must use a local service-account keyfile fallback.
# GOOGLE_APPLICATION_CREDENTIALS=/path/to/service-account.json
#
# Used for: GCS storage, Google Docs API, and BigQuery.
# Required scopes for Google Docs:
#   - https://www.googleapis.com/auth/documents
#   - https://www.googleapis.com/auth/drive
GCS_BUCKET_NAME=your_gcs_bucket_name

# Google Docs Integration (optional)
# If not set, a new document will be created with title "Economic Analysis Reports"
# If you want to use an existing document, set GOOGLE_DOCS_DOCUMENT_ID to the document ID
# (the document ID is the part of the URL after /d/ in Google Docs)
# Example: https://docs.google.com/document/d/DOCUMENT_ID_HERE/edit
# GOOGLE_DOCS_DOCUMENT_ID=your_document_id_here

# Google Drive Integration
# Main folder ID for Google Drive resource (used by gdrive sensor and assets)
# Get the folder ID from the Google Drive folder URL:
# Example: https://drive.google.com/drive/folders/FOLDER_ID_HERE
GOOGLE_DRIVE_FOLDER_ID=your_google_drive_folder_id_here

# Realtor.com Data - Folder ID for Realtor.com CSV files
# The folder should contain subfolders: country/, state/, metro/, county/, zip/
# Each subfolder should contain the corresponding Realtor.com CSV files
# Note: This can be the same as GOOGLE_DRIVE_FOLDER_ID if using one folder
REALTOR_GDRIVE_FOLDER_ID=your_realtor_folder_id_here

# BigQuery auth is through ADC. No service-account keyfile is required when
# local ADC, attached service accounts, or workload identity federation are configured.

# =============================================================================
# DBT
# =============================================================================
DBT_TARGET=dev  # dev, local, or prod

# dbt Platform orchestration from Dagster
# Leave DBT_ORCHESTRATION_MODE=local to run dbt through DbtCliResource.
# Set to dbt_platform after the dbt Platform project/environment is configured.
DBT_ORCHESTRATION_MODE=local
DBT_PLATFORM_ACCOUNT_ID=your-dbt-platform-account-id
DBT_PLATFORM_PROJECT_ID=your-dbt-platform-project-id
DBT_PLATFORM_ENVIRONMENT_ID=your-dbt-platform-environment-id
DBT_PLATFORM_TOKEN=your-dbt-platform-service-token
DBT_PLATFORM_ACCESS_URL=https://cloud.getdbt.com
DBT_PLATFORM_ADHOC_JOB_NAME=economic-data-dagster-adhoc
DBT_PLATFORM_POLL_INTERVAL_SECONDS=60
# Set true in production when dbt Platform owns scheduled builds and Dagster
# should only observe completed dbt Platform runs.
DBT_PLATFORM_OBSERVE_ONLY=false
# DBT_PROJECT_DIR=/path/to/dbt_project  # Optional, auto-detected

# =============================================================================
# CLAUDE CODE / MCP SERVERS
# =============================================================================
# MotherDuck MCP is no longer used — queries now go to BigQuery via the
# dbt-index MCP server's `warehouse` tool (see .mcp.json) or the BigQuery console.

# Claude API - For context documentation maintainer skill
# Get your key from: https://console.anthropic.com/
ANTHROPIC_API_KEY=your_anthropic_api_key_here

# =============================================================================
# GITHUB INTEGRATION (for asset failure monitoring)
# =============================================================================
# GitHub personal access token for creating issues automatically
# Required scopes: repo (full control)
# Get your token from: https://github.com/settings/tokens
GITHUB_TOKEN=ghp_your_personal_access_token_here
GITHUB_REPO_OWNER=C00ldudeNoonan
GITHUB_REPO_NAME=economic-data-project-full

# =============================================================================
# ECONOMIC ALERTS (Gmail API via OAuth user credentials, see #61)
# =============================================================================
# This flow uses a stored refresh token to send mail as a single Gmail
# user — works for personal @gmail.com accounts (where domain-wide
# delegation is not allowed).
#
# One-time setup:
#   1. GCP > APIs & Services > enable Gmail API.
#   2. GCP > OAuth consent screen: External, add yourself as test user,
#      add scope https://www.googleapis.com/auth/gmail.send.
#   3. GCP > Credentials > OAuth client ID (type: Desktop app).
#      Download client_secret.json.
#   4. Run:
#      uv run --project macro_agents python scripts/gmail_oauth_bootstrap.py \
#          --client-secrets /path/to/client_secret.json \
#          --token-output /path/to/gmail_token.json
#   5. Set GMAIL_OAUTH_TOKEN_PATH to the resulting token.json.

# Gmail address that owns the refresh token (also the From: address).
GMAIL_SENDER=you@gmail.com
# Path to the refresh-token JSON written by the bootstrap script.
GMAIL_OAUTH_TOKEN_PATH=/path/to/gmail_token.json
# Recipient inbox for alert emails (same as sender for self-alerts).
ALERT_RECIPIENT=you@gmail.com
DBT_CLOUD_API_TOKEN=
DBT_CLOUD_ACCOUNT=
DBT_CLOUD_PROJECT_ID=
DBT_CLOUD_ENVIRONMENT_ID=